Behind the Data: The Doomsday Clock

In January, I published a piece for the Guardian Datablog about the Doomsday clock being moved closer towards midnight. The clock is a check on how close the world is towards destruction – not literally a clock, but rather a way of analysing the political, environmental and military factors that could start a chain of events that would lead the supposed destruction of the world, at least in part or in whole.

I wrote the article The Guardian, and published the interactive graphic to show how the clock has changed over time:

The process

The challenge with any data is to put it into context. With this, the Bulletin of Atomic Scientists have a website that lists all the times the clock has changed, and what they Bulletin said about the change at the time. This was helpful to provide context to the story; As the chart shows, there have only been four instances where the time to midnight, or total destruction, was closer, and one where it was equal to the five minutes to midnight. Equally, the time in the past few years has gone back down to levels previously seen in the early 50s and the mid 80s, from the time after the end of the Cold War and the end of the Cuban missile crisis, where the clock reached a height of 17 minutes away from midnight in 1991.

Very quickly, and in a small piece of the page, that information can be imparted. The dates can also be compared with whatever the read wants, therefore justifying and making the interactive chart an excellent idea.

The chart itself is a Google trends chart from the Google Docs spreadsheet, and features a scrollable timeline with the years, an expanded timeline to show the changes with more detail, and the explanation data on the side of the chart, so the reader can click on it and find out more information.

To get the information for the dates, I needed to find out when the clock was published for the Google chart to work properly – it wouldn’t display the data in simple year form. To this end, I utilised a Google Books search to find the front covers of the magazine over the years. This was quicker (at 10am in the morning) to find the data than to contact the bulletin direct and ask them to provide me with the exact date, who are based in Chicago, Illinois – a 6 hour time difference away.

Google books would be just as accurate, in any case, as the original front cover with publication date was shown. 15 minutes (if that) of scrolling through and noting the publication date was preferable to the 5 hours of waiting for the bulletin to open, then a few hours on top of that for them to gather the info I needed, call me and email it through.

Further to the Google Doc requiring a date that wasn’t just a year – it also required a date, not just a month. In that instance, I substituted the first date of the month, which put the date in the correct format, and allowed the chart to work.

The spreadsheet behind it ended up looking like this:

The explanation, in the third column, was visible on the final chart, the date in the correct format helped to make the chart work, and and the minutes to midnight was a figure to scroll. Manipulating the data and bringing different sources together worked.

At the end of it I was left with a chart and story with, as of today, 182 tweet shares, 143 Facebook shares, 14 comments generally positive, and an email from the bulletin praising the way it was written and how well I understood the data.